Academic Profile : Faculty

mugshot-tatjencham-400x514.jpg picture
Assoc Prof Cham Tat Jen
Associate Professor, School of Computer Science and Engineering
Assistant Chair (Graduate Studies), School of Computer Science and Engineering (SCSE)
External Links
Tat-Jen is an Associate Professor in the School of Computer Science & Engineering, Nanyang Technological University. He received his BA in Engineering in 1993 and his PhD in 1996, both from the University of Cambridge. Tat-Jen was subsequently conferred a Jesus College Research Fellowship in Science in 1996-97. From 1998 to 2001, he was a research scientist at DEC/Compaq Research Lab in Cambridge, MA, USA. After joining NTU in 2002, he was concurrently a Faculty Fellow in the Singapore-MIT Alliance Computer Science Program in 2003-2006. Tat-Jen has also served a number of terms on the NTU Senate.

In research, Tat-Jen received overall best paper prizes at PROCAMS’05, BMVC’94, and in particular at ECCV'96. Tat-Jen is an inventor on eight patents. He is currently a PI and Fellow of the Rehabilitation Research Institute of Singapore (RRIS), and a PI in the Singtel Cognitive & AI Lab (SCALE@NTU). Previous roles held include being the Director for the Centre for Multimedia & Network Technology (CeMNet) and an affiliate faculty with the Singapore-ETH Centre’s Future Cities Lab. He was also a founding Principal Investigator in the NRF BeingThere Centre (BTC) on 3D Telepresence, a collaboration between NTU, UNC at Chapel Hill and ETH Zurich, which later evolved into the BeingTogether Centre.

Tat-Jen is an Area Chair for CVPR'24, ECCV'24 and ICCV'23, and also an Associate Editor for IEEE T-MM and CVIU. He has previously served as an editorial board member for the International Journal of Computer Vision (IJCV), a General Chair for ACCV’14, as well as senior roles in multiple past premier conferences. Nationally, he has been on various review panels for A*STAR and the National Research Foundation (NRF).
Tat-Jen’s research interests are broadly in computer vision and machine learning, with a current focus on deep learning methods that can exploit semantic and contextual cues. One major thrust is in scene understanding, in particular focusing on high-fidelity 3D reconstruction and decomposition of indoor environments which involves hidden shape prediction / amodal segmentation and scene inpainting / completion, in order to overcome severe clutter and occlusion under limited sensor coverage. Another major thrust is in 3D human analysis, covering diverse topics ranging from face reconstruction and relighting, hand and body pose estimation, to human motion analysis, prediction and ability modeling. Tat-Jen's previous work includes image-based localization in urban environments and projector-camera systems that can turn surfaces into ubiquitous interactive displays.
  • Assistive Robotics Programme